I’ve looked at these types of things in the past, and I think they are fundamentally flawed. As a reader you can dip into the start of a story, choose the branches that interest you personally, but rapidly come to a section that still needs writing. It’s a frustrating experience as a user, even ignoring the quality of some of the writing, especially for erotic writing - talk about orgasm denial!
It’s easy to see why: if each section offers two choices, then you’ve got combinatorial explosion - go 10 choices in and there are 1,024 (2^10) pages that still need writing. Any many pages offer even more choices than two (3^10 is a much bigger number, ~60x bigger, than 2^10)! Even with an army of contributors that’s not going to get done. The only way around it is to be able to join threads back together again, but few actually support this, and you also need variables to track what has happened (at least CHYOA offers some, but only booleans and ints), and lots of conditions to know what to offer next.
At a bare minimum for a WG story you need to be able to track the size of the player and each NPC, and have some kind of ‘widget’ that can substitute their current size into any description in a repeatable way, and a policy of how the writers use the variables. And that means the writers have to be at least passingly familiar with some kind of code - or you have a writer and coder working in tandem, and someone to coordinate the approach (which starts to break the idea of public submission).
Which kind of brings me onto what does work and achieves some kind of completion; think of games that do have many scenes submitted by many authors - it’s teams of people working in a coordinated way, there’s a plan, an overarching story and world. I guess CoC and TiTS are good examples (Pornarium also, back in the day). They are/were to some extent open to public submission, but they were also curated, and had something more complex than just hyperlinks at their core.
What might work is a combination of Twine, Tweego, and Git. Many people can work in Twine, it’s not too complex, use Tweego for the build and authors can work their parts of the story in separate files. Version control in Git (though not simple) would allow anyone to take the story and add to it as they will, but also allow changes to be merged back into the master (or not).
Possibly inklewriter, but I don’t know if you can share logins at all and how it might handle simultaneous editing?
In short, I think the original concept is an intractable problem. Public submission and no curation is at odds with producing CYOA style stories that are finished from the reader’s perspective (while there may be some combinations of choices that are complete, the readers chance of finding them is vanishingly small). You have to give up one of those things: make it not public, add a curator, or just accept readers are going to end up disappointed.
As a writer, I assume the goal is first to have people read your story, and collaborating to some extent is a secondary concern!
Another issue I have with the likes of writing.com is they way some seem to treat it essentially a virtual begging bowl: I want a story about ‘x’ so I write the start and see if someone else would like to “finish” (ie. do all the work) for me.
As for the rules that apply on these sites, I suspect it’s a basic CYA exercise (much like we have here) as hosting illegal content is, well, illegal - obviously what exactly is illegal depends on where the hosting is done, and the jurisdiction applying to the owners. From that point of view only CHYOA has a sensible approach!